The context

Lauren ran operations for product and technology at a PE-backed fintech with about 5,000 employees. She was also co-leading the company’s AI transformation, with value targets already committed to the board.

Then a generous vendor deal expired, and the real cost of AI showed up all at once. Newer models, broad adoption, and no usage guardrails had pushed spend far above plan. Pulling back premium models created friction with employees who’d gotten used to them, and budget season was a month away.

The advice

Lauren talked it through with three product and technology executives, including an SVP of Product and a former product leader at a national retailer, all working through AI costs at their own companies.

• Set your own narrative. Don’t let the board’s cost-cutting frame define AI spend. Position it as an enterprise software investment with reusable tools.
• Find where the spend is concentrated. A small group of heavy users usually drives most of the cost. Ask them to show the return, and let high-value teams earn access to premium models.
• Make cost visible. Most employees have no idea what frontier models cost, and transparency alone changes behavior.
• Learn from a real pilot. One coach’s company gave a fixed-budget pilot to about 20% of employees, with a clear application process tied to specific use cases.
• Measure outcomes, not usage. Are you getting the same or better results at the same or better cost?

The outcome

The idea that clicked for Lauren was simple: model AI as a new per-employee cost line. It turned a sprawling, scary number into something she could plan, forecast, and allocate to business units heading into budget season. She also left with a real benchmark to work from and an intro to go deeper on AI across the product development lifecycle.

Why it worked

Every coach was managing the same problem live, so the advice came with real numbers.

The group moved her from defending costs to owning the narrative.

The per-employee frame connected AI directly to headcount planning, which leadership already understood.

The broader point

Some problems are too new for anyone to have a playbook. The fastest way through is a room of peers solving the same thing at the same time.

The advice might seem obvious now. Having peers to act on it with is what made the difference.

Don’t let the board’s cost-cutting frame define AI spend. Set your own narrative.